Came across this, the "Gentoo Linux Install Script". It's an dialog based console install script for Gentoo that aims to both take the pain out of installing Gentoo as well as letting you do all of the configuration at the beginning of the install, so that you can then leave it to install unwatched. I downloaded it, gave it a quick looking over and e-mailed the author with some usability ideas. He was very open to input and said he will use my ideas when he gets to improving the UI.
I've been hoping for a way to make the gentoo install more sane. It's not likely to be as easy as Redhat(TM) for quite a while, if ever, but using glis would make it easy enough for any windows power user. I think that this is a great compromise between the gtk2 install of redhat(using X on the install cd creates problems) and the needlessly hard install that Gentoo has atm.
This fits in quite well with my plans for a gnome distro. I could make a custom stage 3 livecd with glis as the installer and you could be up and running with the gnome linux desktop in <30mins. Using gentoo has the benefit that I can create a default package selection but the user will have a massive collection of software at their findertips - so they will get both the sensible "Just Works" environment by default, but if they are that way enclined they can partake the massive collection of software avaliable. Also, from a development point of view - it's much easier to write ebuilds that to produce rpm, I just need to find out how I go about making a customised stage3 tarball.
Stage3 is a tarball of enough binary packages needed to make up a full graphical system. It makes gentoo quicker to install but you loose out on some of the optomisations. But you can still recompile everything later _after_ you have a running system, so I think that stage3 install is the way to go.
I think it's a pity that Zynot wont limit itself to embedded Linux and leave x86 to Gentoo. I have plans to do some work with the Zynot gnome/gpe developer, but I think that both he and I still think that Gentoo is where it is at for the desktop.
Distributed Usability Testing
This is something that I have been thinking about for quite a while. Usability testing costs money and Gnome is lucky enough to have sun investing in it's usability, but there is still a lot that can be done by us, the members of OSS projects.
I started thinking about distributed usability testing when my girlfriend started using my pc to check her e-mails. I was using Galeon as my browser at the time(epiphany now) and I filed 2 or 3 bug reports based on problems that she had. Issues that had not really occured to me but that made a lot of sense when she pointed them out.
So what is Distributed Usability Testing? it's testing carried out in a casual ongoing way. All of us have windows using friends and they are perfect to try out OSS software on. There are two way to do this IMHO:
1. If you have a friend/family memeber who understands how important OSS is to you and is willing to put up with doing some actual tasks then you can use them to perform an actual test, the way they are done in usability labs, except on a smaller scale. According to the statistics I think you need to have 5 or 6 people doing a usability test to find most of the problems. So if you can find 6 hackers/users with a friend who is willing to help then you have yourself a free usability test.
2. This way is more sneaky and less regemented. It involves doing usability tests on people who don't know they are being tested. They simplest example is the above story about my girlfriend: you just sit someone down infront of your opensource pc and let them try to use it. Make mental notes of the problems they encounter. Of course it is not always possible to bring people to you pc - so use a live cd like knopix. Tell them you want to show them Linux and see how well it works for them. I'm not suggesting you be blatently dishonest, a lot of the time you can just tell the person what you are doing. On the upside you are mixing Linux advogacy with Distributed Usability testing.
To fully capitalise on this Idea it would be necessary for some usability engineers (eg. the Gnome Usability Project) to make a guide to Distributed Usability testing and to regularly publish usability tests for use by the testers.
Another idea is to join forces with the soon to be created Gnome Marketing project and produce a Gnome Livecd suitable for increaing Gnome awareness as well as doing Distributed Usability testing. I'm going away tomo, but when I get back I may propose this mental idea to the Gnome Usability list.
Must get a better blog going next year - one that I can e-mail to like Jeff's.